English

• Recent years have witnessed the growing applications of ketoreductase (KRED)-catalyzed dynamic reductive kinetic resolution (DYRKR) to the asymmetric synthesis of chiral alcohols containing two adjacent stereogenic centers [1–4]. In addition to the intrinsic merits of biocatalysis including sustainability and mild operation conditions [5,6], undoubtedly this increasing popularity is mainly ascribed to the capability of predominantly forming one of the four possible stereoisomeric products in one-step, with a maximum theoretical yield of 100% [1–4]. Within this context, our group very recently performed the first systematic synthesis of various anti-aryl-(2S,3S)-2–chloro-3–hydroxy esters (anti-(2S,3S)-1, Fig. 1) through ketoreductase LfSDR1-catalyzed DYRKR of racemic aryl α–chloro β-keto esters (2), and further realized the chemo-enzymatic synthesis of the calcium channel blocker d-cis-diltiazem as wells as its structural analogs clentiazem and siratiazem [7].

  Figure 1

  

  Figure 1.  Structures of anti-aryl-(2S,3S)-2–chloro-3–hydroxy esters (anti-(2S,3S)-1), aryl α–chloro β-keto esters (2), syn-aryl-(2S,3R)-2–chloro-3–hydroxy esters (syn-(2S,3R)-1), N-benzoyl-2R, 3S-3-phenyl isoserine (3), TA-993 (4), and l-cis-diltiazem (5).

  DownLoad: Full-Size Img PowerPoint

  Given the fact that not only anti-(2S,3S)-1, but also their diastereomers, syn-aryl-(2S,3R)-2–chloro-3–hydroxy esters (syn-(2S,3R)-1, Fig. 1) are valuable for the synthesis of pharmacologically important molecules, such as N-benzoyl-2R, 3S-3-phenyl isoserine (3), the critical structure component of taxol [8], antiplatelet agent TA-993 (4) [9,10], and the cyclic nucleotide gated channel blocker l-cis-diltiazem (5) [11–13], it is desirable to develop a generic, stereoselective synthesis of syn-(2S,3R)-1 via KRED-catalyzed DYRKR. To the best of our knowledge, the only relevant studies were ketoreductases YDL124w-, and CaADH-catalyzed highly stereoselective synthesis of syn-(2S,3R)-1a (Scheme 1), the synthetic precursor of 3 [14,15]. Notably, chemical reduction of racemic aryl α–chloro β-keto esters (2) catalyzed by chiral ruthenium complexes only furnished syn-(2S,3R)-1 with poor-to-moderate diastereoselectivities (< 9:1 dr) [16–18]. Herein, we report the enzyme identification, directed evolution, and application of the evolved variant to the asymmetric synthesis of a variety of syn-aryl-(2S,3R)-2–chloro-3–hydroxy esters via DYRKR.

  Scheme 1

  

  Scheme 1.  Stereoselective synthesis of syn-(2S,3R)-1 through ketoreductase-catalyzed DYRKR of aryl α–chloro β-keto esters (2).

  DownLoad: Full-Size Img PowerPoint

  In our previous study [7], we found that three ketoreductases, namely KRED-F42, KdoADH, and KRED-Bt could reduce the model substrate 2b to furnish syn-(2S,3R)-1b as the major product (Table 1, entries 1–3). Nevertheless, the diastereoselectivities were still unsatisfactory (1.1:1–6.2:1 dr). Prompted by the report that syn-(2S,3R)-1a was generated in a highly stereoselective manner through YDL124w-catalyzed reduction of 2a (Scheme 1) [14], we tested this enzyme for the reduction of 2b. To our delight, the target (2S,3R)-1b was afforded in 98.4% conversion with 46:1 dr and 98.1% ee (Table 1, entry 4). The relative syn-configuration of thus formed (2S,3R)-1b was determined based on the J_{2, 3} value of 6.7 Hz, whereas the absolute configuration was initially assigned according to the optical rotation data and further unambiguously confirmed by X-ray crystallography (Fig. S1 in Supporting information). Next, we purified the C-terminal-His₆-tagged YDL124w and determined its specific activity towards 2b as 0.086 U/mg (Table 2). Given the moderate catalytic activity, we decided to engineer the enzyme first prior to examining the substrate scope of this biocatalytic reaction [19–21].

  Table 1

  

  Table 1.  KRED-catalyzed stereoselective reduction of α–chloro β-keto ester 2b.^a

  DownLoad: CSV

  Table 2

  

  Table 2.  Specific activity of the purified YDL124w and its evolved variant YDL124w-S99T/F225Y toward substrates 2a, 2b, and 2c.^a

  DownLoad: CSV

  As a crystal structure of YDL124w is not available, we opted to invoke the error-prone PCR (epPCR) approach. Specifically, by employing the wild-type (WT) YDL124w as the template, epPCR was conducted with 0.3 mmol/L Mn²⁺ resulting in an average mutation rate of 1–3 amino acid substitutions. Upon screening roughly 36, 000 colonies using activity assays performed with cell-free extracts of mutant enzymes, an enzyme variant with about 20-fold increased activity towards substrate 2b was discovered. Sequence determination indicated this variant carried two mutations: S99T/F225Y. Hence, it was named as YDL124w-S99T/F225Y. The effect of different pH conditions ranging from 5.0 to 8.0, on the specific activity of the purified YDL124w-S99T/F225Y was examined, with pH 6.5 being identified as the optimal one (Fig. S2 in Supporting information). Within the temperatures between 20 ℃ and 50 ℃ tested, YDL124w-S99T/F225Y showed the highest specific activity at 30 ℃ (Fig. S3 in Supporting information). Intriguingly, assays on purified YDL124w-S99T/F225Y under optimal conditions suggested this enzyme only possessed 1.9-fold higher specific activity towards 2b relative to WT (Table 2). Furthermore, we also compared the soluble expression levels of WT and YDL124w-S99T/F225Y using the SDS-page analysis. As seen in Fig. S4 (Supporting information), this evolved enzyme variant was better solubly expressed than WT. Therefore, both the improved specific activity and the enhanced expression likely contributed to the aforementioned 20-fold increased activity in the lysate-based assays. In addition to 2b, we also determined the specific activity of WT and YDL124w-S99T/F225Y toward another two α–chloro β-keto esters 2a and 2c, because the bioreduction products of them, namely syn-(2S,3R)-1a and syn-(2S,3R)-1c, are critical synthetic intermediates to N-benzoyl-2R, 3S-3-phenyl isoserine (3) and TA-993 (4), respectively. Compared with WT, YDL124w-S99T/F225Y exhibited 1.4-fold elevated specific activity towards both 2a and 2c (Table 2).

  With an improved enzyme variant YDL124w-S99T/F225Y in hand, 17 chemically-prepared aryl α–chloro β-keto esters 2 were deployed for examining its synthesis versatility at a preparative-scale (0.6 mmol) (Scheme 2). Firstly, the current bioreduction of substrates equipped with a para-substituent at the phenyl ring (2b, 2c, 2e-2j) or substrate without a substituent (2d), all occurred with high levels of stereoselectivity, regardless of the substituent's electronic property. The corresponding products syn-(2S,3R)-1 were isolated in 88%−98% yield along with 19:1–66:1 dr and 89%~ > 99% ee. In comparison, anti-(2S,3S)-1f, anti-(2S, 3S)-1g, and anti-(2S,3S)-1h, containing electron-withdrawing groups (F, Cl, Br), were synthesized with only moderate diastereoselectivities (5.3:1–9.2:1 dr) through ketoreductase LfSDR1-catalyzed DYRKR [7]. Secondly, sterically demanding substrates, including 2k and 2n with a para-phenyl and an ortho–methoxy substituent, respectively, the naphthyl-based 2l, and the bis–methoxy groups containing 2o were all reduced smoothly by YDL124w-S99T/F225Y, providing the desired products (syn-(2S,3R)-1k, syn-(2S,3R)-1n, syn-(2S,3R)-1l, and syn-(2S,3R)-1o) in good to excellent yields with excellent stereoselectivities (73%−96% yield, 20:1–83:1 dr, 95%~ > 99% ee). Thirdly, not only methyl ester, but ethyl ester 2a was accepted by YDL124w-S99T/F225Y, furnishing syn-(2S,3R)-1a in 99% yield with 44:1 dr and 94% ee. Finally, although ketone 2m with a meta–methoxy-substituted phenyl ring and heteroaryl-based substrates (2p and 2q) could be transformed to the respective chlorohydrins syn-(2S,3R)-1m, syn-(2S,3R)-1p, and syn-(2S,3R)-1q in 82%−91% yield, the diastereoselectivities were inadequate (6.1:1–10:1 dr). In the future, reaction optimization and/or protein engineering might be utilized to boost the stereoselectivity.

  Scheme 2

  

  Scheme 2.  Preparative-scale synthesis of syn-(2S,3R)-1 catalyzed by YDL124w-S99T/F225Y. Reaction conditions (60 mL): 2 (10 mmol/L), glucose (15 mmol/L), NADP⁺ (0.2 mmol/L), 10 mL of 15% (w/v) CFE of YDL124w-S99T/F225Y in NaP_i buffer (100 mmol/L, pH 6.5), 0.2 mL of 15% (w/v) CFE of GDH in NaP_i buffer (100 mmol/L, pH 6.5), and 6 mL of MeOH, in 44 mL of NaP_i buffer (100 mmol/L, pH 6.5). Reaction mixtures were incubated at 30 ℃ with 520 rpm stirring for 12 h. Isolated yield was given. The dr values were determined by the ¹H NMR analysis of the crude product. The ee values were determined by chiral HPLC analysis. The racemic product standards were obtained through NaBH₄-mediated reduction of 2.

  DownLoad: Full-Size Img PowerPoint

  The application potential of the currently developed DYRKR method was further showcased by the reduction of 2b at an industrially useful concentration of 120 g/L. In our previous study of LfSDR1-catalyzed reduction of 100 g/L of 2b, a continuous fed-batch operation was necessary in order to achieve high reaction conversion, as LfSDR1 was severely inhibited by 2b at substrate concentration of 40 mmol/L or greater [7]. Fortunately, no obvious substrate inhibition was observed for YDL124w-S99T/F225Y with up to 80 mmol/L of 2b. Hence, a single-dose of 2b was employed for the current study. In practice, 2b (120 g, 120 g/L), glucose (178.5 g, 2.0 equiv.), NADP⁺ (149 mg, 0.2 mmol/L), methanol (100 mL), cell-free extracts of YDL124w-S99T/F225Y (corresponding to 90 g/L of wet cells), and cell-free extracts of GDH (corresponding to 3.3 g/L of wet cells) in NaP_i buffer (900 mL, 100 mmol/L, pH 6.5) were stirred at 30 ℃, while maintaining the reaction pH between 6.0 and 6.5 by titrating a 4 mol/L NaOH solution. After the reaction went to nearly completion (98% conversion) as indicated by the chiral HPLC analysis, celite was added, followed by filtration. The filter cake was washed three times with ethyl acetate, and the two layers of filtrate were separated. The aqueous solution was extracted further with ethyl acetate for 2 times. Finally, the combined organic layer was washed with brine, dried, and concentrated in vacuo to give syn-(2S,3R)-1b (110.6 g, 91.4% yield) with excellent purity (93.6% chemical purity, 41:1 dr, and 98.4% ee).

  Molecular docking was adopted to shed light on the molecular basis for the increased catalytic activity of YDL124w-S99T/F225Y. First of all, 3-D structure models of YDL124w and YDL124w-S99T/F225Y were built with SWISS-MODEL server [22], by using the crystal structure of the conjugated polyketone reductase C2 (CPR-C2) from Candida parapsilosis IFO 0708 as the template (PBD ID: 4H8N) [23]. The amino acid sequence identity between the wild-type YDL124w and CPR-C2 is 47%. Both models thus generated were considered as reasonable, based on the respective VERIFY [24] values of 88.8% and 88.5%, ERRAT [25] values of 92.3 and 90.9, as well as both 100% of the residues being found in the allowed region of the Ramachandran plots [26]. Then, substrate 2b was docked into the modeled structures of wild-type of YDL124w and variant YDL124w-S99T/F225Y, respectively (Fig. 2). The distance between the hydride of C4 from the nicotinamide of NADPH and the carbon atom of the carbonyl group of 2b, as well as the distance between the oxygen atom of the carbonyl group of 2b and the phenoxyl hydrogen of the catalytic residue Tyr64, were both shorter for YDL124w-S99T/F225Y than those for the wild-type enzyme (4.9 Å and 4.6 Å versus 6.1 Å and 6.2 Å). These results indicated that compared with WT, YDL124w-S99T/F225Y had higher chances to conduct both events of the hydride and the proton transfer from NADPH and Tyr64, respectively, to the carbonyl group of 2b, hence, in accordance with the improved catalytic activity of this variant determined experimentally. Finally, the two mutated residues, Ser99 and Phe225 were found to be located distal from the active site. Although at this moment it is unclear how these mutations exactly affect the catalytic activity, one possibility is that nonlocal structural perturbations occurred after mutation, thereby influencing the catalytic activity indirectly through changing the enzyme's conformational motions as well as the probability of sampling conformations conducive to catalysis [27].

  Figure 2

  

  Figure 2.  Docking of substrate 2b into the active sites of wild-type of YDL124w (a) and variant YDL124w-S99T/F225Y (b).

  DownLoad: Full-Size Img PowerPoint

  In summary, epPCR-based directed evolution of ketoreductase YDL124w enabled the development of an enzyme variant, YDL124w-S99T/F225Y, exhibiting 1.4–1.9 fold enhancement of specific activity relative to the wild-type enzyme towards α–chloro β-keto esters 2a, 2b, and 2c. Compared with YDL124w, the soluble expression level of YDL124w-S99T/F225Y was much improved as well. Seventeen synthetically useful syn-aryl-(2S,3R)-2–chloro-3–hydroxy esters (syn-(2S,3R)-1) were synthesized through YDL124w-S99T/F225Y-catalyzed dynamic reductive kinetic resolution of racemic aryl α–chloro β-keto esters (2) in 73%−99% isolated yields, along with 6.1:1–83:1 dr and 31%~ > 99% ee. The practical synthesis potential of the method developed in this study was demonstrated by a nearly complete reduction of 120 g/L of 2b at a hectogram scale, furnishing syn-(2S,3R)-1b, a synthetic intermediate to the cyclic nucleotide gated channel blocker l-cis-diltiazem (5), in 91.4% isolated yield with 41:1 dr and 98.4% ee. Molecular docking was performed to help provide insights for the improvement of catalytic activity of the evolved enzyme variant.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  The National Key Research and Development Program of China (Nos. 2021YFA0911400 and 2021YFF0600704), and the National Natural Science Foundation of China (Nos. 22071033 and 21801047) are acknowledged for the financial supports.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108178.

  
  
• 1. [1]

     G.A. Applegatea, D.B. Berkowitz, Adv. Synth. Catal. 357 (2015) 1619–1632. doi: 10.1002/adsc.201500316

  2. [2]

     Z. Li, H. Yang, J. Liu, Z. Huang, F. Chen, Chem. Rec. 21 (2017) 1611–1630.

  3. [3]

     G. Ni, J. Tang, J. Zou, et al., Chin. J. Org. Chem. 39 (2019) 339–349. doi: 10.6023/cjoc201806012

  4. [4]

     C. Huang, J. Liu, J. Fang, et al., Front. Bioeng. Biotechnol. 10 (2022) 929784. doi: 10.3389/fbioe.2022.929784

  5. [5]

     S. Wu, R. Snajdrova, J.C. Moore, K. Baldenius, U.T. Bornscheuer, Angew. Chem. Int. Ed. 60 (2021) 88–119. doi: 10.1002/anie.202006648

  6. [6]

     R.A.A. Sheldon, D. Brady, M.L.L. Bode, Chem. Sci. 11 (2020) 2587–2605. doi: 10.1039/C9SC05746C

  7. [7]

     X. Yue, Y. Li, M. Liu, et al., Chem. Commun. 58 (2022) 9010–9013. doi: 10.1039/D2CC03102G

  8. [8]

     I. Rimoldi, M. Pellizzoni, G. Facchetti, et al., Tetrahedron 22 (2011) 2110–2116. doi: 10.1016/j.tetasy.2011.11.017

  9. [9]

     M. Katoh, T. Karasawa, A. Watanabe, et al., Biol. Pharm. Bull. 24 (2001) 501–504. doi: 10.1248/bpb.24.501

  10. [10]

      H. Narita, M. Kaburaki, H. Doi, A. Yasoshima, S. Murata, Jpn. J. Pharmacol. 68 (1995) 397–404. doi: 10.1254/jjp.68.397

  11. [11]

      J.C. Gomora, J.J. Enyeart, J. Pharmacol. Exp. Ther. 290 (1999) 266–275.

  12. [12]

      L.W. Haynes, J. Gen. Physiol. 100 (1992) 783–801. doi: 10.1085/jgp.100.5.783

  13. [13]

      R. Shalom, L. Barki-Harrington, G. Rimon, Eur. J. Pharmacol. 543 (2006) 8–13. doi: 10.1016/j.ejphar.2006.06.018

  14. [14]

      B.D. Feske, I.A. Kaluzna, J.D. Stewart, J. Org. Chem. 70 (2005) 9654–9657. doi: 10.1021/jo0516077

  15. [15]

      G.A. Applegate, R.W. Cheloha, D.L. Nelson, D.B. Berkowitz, Chem. Commun. 47 (2011) 2420–2422. doi: 10.1039/C0CC04585C

  16. [16]

      J. Bai, S. Miao, Y. Wu, Y. Zhang, Chin. J. Chem. 29 (2011) 2476–2480. doi: 10.1002/cjoc.201180419

  17. [17]

      L. Qiu, F.Y. Kwong, J. Wu, et al., J. Am. Chem. Soc. 128 (2006) 5955–5965. doi: 10.1021/ja0602694

  18. [18]

      L. Qiu, J. Qi, C.C. Pai, et al., Org. Lett. 4 (2002) 4599–4602. doi: 10.1021/ol026817+

  19. [19]

      K. Chen, F.H. Arnold, Nat. Catal. 3 (2020) 203–213. doi: 10.1038/s41929-019-0385-5

  20. [20]

      G. Qu, A. Li, C.G. Acevedo-Rocha, Z. Sun, M.T. Reetz, Angew. Chem. Int. Ed. 59 (2020) 13204–13231. doi: 10.1002/anie.201901491

  21. [21]

      J. Li, G. Qu, N. Shang, et al., Green Synth. Catal. 2 (2021) 45–53. doi: 10.1016/j.gresc.2021.01.005

  22. [22]

      A. Waterhouse, M. Bertoni, S. Bienert, et al., Nucleic Acids Res. 46 (2018) W296–W303. doi: 10.1093/nar/gky427

  23. [23]

      H.M. Qin, A. Yamamura, T. Miyakawa, et al., Appl. Microbiol. Biotechnol. 98 (2014) 243–249. doi: 10.1007/s00253-013-5073-9

  24. [24]

      D. Eisenberg, R. Lüthy, J.U. Bowie, Methods Enzymol. 277 (1997) 396–404.

  25. [25]

      C. Colovos, T.O. Yeates, Protein Sci. 2 (1993) 1511–1519. doi: 10.1002/pro.5560020916

  26. [26]

      S.C. Lovell, I.W. Davis, W.B. Adrendall, et al., Proteins 50 (2003) 437–450. doi: 10.1002/prot.10286

  27. [27]

      K.F. Wong, T. Seizer, S.J. Benkovic, S. Hammes-Schiffer, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 6807–6812. doi: 10.1073/pnas.0408343102

• 

  Figure 1  Structures of anti-aryl-(2S,3S)-2–chloro-3–hydroxy esters (anti-(2S,3S)-1), aryl α–chloro β-keto esters (2), syn-aryl-(2S,3R)-2–chloro-3–hydroxy esters (syn-(2S,3R)-1), N-benzoyl-2R, 3S-3-phenyl isoserine (3), TA-993 (4), and l-cis-diltiazem (5).

  下载: 全尺寸图片 幻灯片
  

  Scheme 1  Stereoselective synthesis of syn-(2S,3R)-1 through ketoreductase-catalyzed DYRKR of aryl α–chloro β-keto esters (2).

  下载: 全尺寸图片 幻灯片
  

  Scheme 2  Preparative-scale synthesis of syn-(2S,3R)-1 catalyzed by YDL124w-S99T/F225Y. Reaction conditions (60 mL): 2 (10 mmol/L), glucose (15 mmol/L), NADP⁺ (0.2 mmol/L), 10 mL of 15% (w/v) CFE of YDL124w-S99T/F225Y in NaP_i buffer (100 mmol/L, pH 6.5), 0.2 mL of 15% (w/v) CFE of GDH in NaP_i buffer (100 mmol/L, pH 6.5), and 6 mL of MeOH, in 44 mL of NaP_i buffer (100 mmol/L, pH 6.5). Reaction mixtures were incubated at 30 ℃ with 520 rpm stirring for 12 h. Isolated yield was given. The dr values were determined by the ¹H NMR analysis of the crude product. The ee values were determined by chiral HPLC analysis. The racemic product standards were obtained through NaBH₄-mediated reduction of 2.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  Docking of substrate 2b into the active sites of wild-type of YDL124w (a) and variant YDL124w-S99T/F225Y (b).

  下载: 全尺寸图片 幻灯片

  Table 1.  KRED-catalyzed stereoselective reduction of α–chloro β-keto ester 2b.^a

  下载: 导出CSV

  Table 2.  Specific activity of the purified YDL124w and its evolved variant YDL124w-S99T/F225Y toward substrates 2a, 2b, and 2c.^a

  下载: 导出CSV
•